Neuritic plaques containing primarily amyloid beta protein (Abeta) are one of the hallmarks of Alzheimer's Disease. Beta-site APP cleaving enzyme (BACE), known also as beta-secretase, Asp2, and Memapsin, has been identified as the enzyme responsible for processing amyloid precursor protein (APP) to produce the N-terminal portion of the Abeta peptide. This enzyme has been suggested as rate limiting in the production of the Abeta peptide. See, for example, Sinha et al., 1999, Nature 402:537-554, and published PCT applications WO 00/17369, WO 01/23533, and WO 98/22597. See also: Hussain, I. et al., 1999, Mol. Cell. Neurosci. 14:419-427; Vassar, R. et al., 1999, Science 286:735-741; Yan, R. et al., 2000, Nature 402:533-537; and Lin, X. et al., 2000, Proc. Natl. Acad. Sci. USA 97:1456-1460 (2000).
BACE is a therapeutic target for the development of inhibitory compounds for the treatment of Alzheimer's Disease. Rational drug design methods require supply of properly expressed, refolded, and active BACE in order to model and design appropriate new drugs. BACE in sufficient amounts has proven difficult to obtain.
BACE is a relatively large and structurally complex enzyme. The primary structure of BACE as it is synthesized in the endoplasmic reticulum is shown in FIG. 1 [SEQ ID NO:1]. The enzyme contains 501 amino acids, including a N-terminal signal (leader) sequence of about 21 amino acids (pre-sequence domain) followed by a pro-sequence domain consisting approximately of residues 22 to 45 (pro-sequence domain) that is proteolytically removed once the enzyme reaches its destination in the Golgi apparatus, to generate a mature enzyme.
Prosequence domains are commonly found in protease precursor polypeptides, where they generally function to prevent catalytic activity and assist in protein folding. The pro-sequence domain is typically cleaved from the protease precursor to generate a mature active protease. Previous work in a baculovirus expression system, expressing a BACE precursor having BACE pre-sequence and pro-sequence domains but truncated at the junction between the putative protease and transmembrane region, indicated that the pro-sequence domain facilitates proper folding of the BACE protease domain. See Shi, X-P. et al., Mar. 30, 2001, J. of Biol. Chem. 276(13):10366-10373.
BACE contains a transmembrane domain of about 27 amino acids that anchors the protein to the membrane. A short cytosolic C-terminal tail of 21 amino acids follows the transmembrane domain. Attachment to the membrane allows BACE to interact with and cleave APP, the first and prerequisite step in the generation of A-beta.
BACE isolated from human brain is heavily glycosylated. As expressed by a stably transfected 293T cell line, BACE is glycosylated at four asparagines: 132, 151, 202, and 333. Analysis of HEK 293 cells stably overexpressing BACE showed that the enzyme is phosphorylated at Ser477, and that phosphorylation regulates enzyme intracellular trafficking (Walter et. al., 2001, J. Biol. Chem. 276:14634-41). Three disulfide bonds suggested as critical for activity, are formed between the following pairs of cysteine residues: Cys195-Cys399, Cys257-Cys422, and Cys309-Cys359 (Haniu et. al., 2000, J. Biol. Chem. 275:21099-21106).
These structural features of the BACE polypeptide all appear to have specific functions relating to enzymatic activity. Enzymes expressed in insect and CHO cells are properly refolded and show activity. These proteins are glycosylated. For example, insect cells express glycosylated BACE, from the mannose-rich glycans available in the insect cells. Biantennary and triantennary oligosaccharides of the complex type provide glycosylation in the CHO-expressed BACE (Charlwood et. al., 2001 J. Biol. Chem. 276:16739-48). These glycosylated proteins have proven difficult to process, however, due to the heterogeneity conferred by differential glycosylation. BACE production in these cells generally requires expensive culture media and does not yield high amounts of protein.
These negative aspects of mammalian and insect cell expression do not apply to BACE proteins expressed in E. coli. Bacteria are easy to grow, produce high yields of protein, and the cost of culture media is low. However, E. coli do not provide for post-translational modifications of the protein.
Previous attempts to produce and isolate large quantities of active BACE from E. coli were initially unsuccessful. Although the cells grew well and expressed reasonable quantities of protein, refolding and isolating the enzyme using known published methods, including those described in Lin et. al. 2000, PNAS USA 97:1456-60 and Tang, WO 01/00663, failed to produce quantities of active BACE suitable for drug discovery methods.
Accordingly, a simple, efficient, and reliable method for expression of recombinant BACE in E. coli and for refolding and purification of sufficient quantities of BACE necessary for use in drug discovery methods is greatly needed.